PSO-based sparse regularization approach for structural damage detection

Abstract

Structural damage detection (SDD) is one crucial step in the field of structural health monitoring (SHM). A novel particle swarm optimization (PSO)-based sparse regularization approach is proposed for the SDD problem in this study. Where, the classical first-order sensitivity analysis and the l 1/2 -norm regularization are introduced to define the objective function, so the SDD problem is converted into a kind of optimization problem which may be solved by the PSO accordingly. The approach includes two steps: the model updating based on sensitivity analysis is employed for the SDD in the first step, and some rough damage locations of structures are obtained by the PSO. Then only the possible damage locations are considered and the PSO is further used to quantify damage extents in the second step for higher SDD accuracy. The illustrated numerical simulations show that the proposed approach can not only effectively locate structural damage but also quantify damage extents with good accuracy as well as strong robustness to noises.